NETCONF keepalive request uses `default-request-timeout-millis` as timeout until which the keepalive RPC reply must be received.

Problem is that `default-request-timeout-millis` is usually set by a user based on longest taking request (like get entire config / operational) which may be several tens of minutes. However, keepalive reply is expected to be received quicker. That means when `keepalive-delay` is 10 seconds and `default-request-timeout-millis` is 30 minutes then keepalive mechanism can detect session is down after 30 minutes soonest.